1
$\begingroup$

I want to evaluate an integral which is essentially a sum of the form $\lim_{a\to 0^+}\int_{-\infty}^\infty \frac{e^{ikx}}{(x+ia)^m} dx$, where $k\in\mathbb R$ and $m\in\mathbb N$. By the residue theorem, we can easily evaluate the integral $$\int_{-\infty}^\infty \frac{e^{ikx}}{(x+ia)^m} dx = -\frac{2\pi i}{(m-1)!} (ik)^{m-1} e^{ka} \theta(-k), \tag{*}$$ where $\theta$ is the Heaviside function. In particular, letting $a\to 0^+$, we simplify $$\lim_{a\to 0^+}\int_{-\infty}^\infty \frac{e^{ikx}}{(x+ia)^m} dx = -\frac{2\pi i}{(m-1)!} (ik)^{m-1} \theta(-k).$$

Now, I have a somewhat complicated expression, which is obtained from other Mathematica calculation:

$Assumptions = {a > 0, k1 > 0, k2 > 0, k3 > 0, k1 + k2 - k3 > 0}
exp = 1/(x3 + I a)^3 E^(-I (k1 + k2 - 
    k3) x3) (-3 I (-1 + E^(I k1 x3)) (-1 + E^(I k2 x3)) k2 + 
   3 k1 (I + I E^(I (k1 + k2) x3) + E^(I k1 x3) (-I + k2 x3) - 
      E^(I k2 x3) (I + k2 x3)))

Question: I want to evaluate

Integrate[exp, {x3, -Infinity, Infinity}]/.{a->0}

Although exp is complicated, it is a sum of the form $(*)$, so it should be calculable. However, Mathematica cannot evaluate this in a reasonable time.

I think I have to somehow "teach" Mathematica to use $(*)$. How can I do this?

$\endgroup$
2
  • 1
    $\begingroup$ This gives a result, but I'm not sure it's what you want: Integrate[Simplify@#, {x3, -Infinity, Infinity}, Assumptions -> {a > 0, k1 > 0, k2 > 0, k3 > 0, k1 + k2 - k3 > 0}] & /@ Expand[exp] $\endgroup$
    – Michael E2
    Feb 20 at 2:34
  • 1
    $\begingroup$ You probably want to rewrite the integral in terms of FourierTransform, though. $\endgroup$
    – Michael E2
    Feb 20 at 2:36

1 Answer 1

5
$\begingroup$

Maybe this?:

Assuming[{a > 0, k1 > 0, k2 > 0, k3 > 0, k1 + k2 - k3 > 0},
 Total@Map[
   With[{e = First@Cases[#, E^_]}, 
     Sqrt[2 Pi] FourierTransform[#/e, x3, k] /. k -> I*Last[e]/x3] &,
   Simplify /@ List @@ Expand[exp]]
 ]
(*
-3 E^(-a k3) k1 k3^2 \[Pi] HeavisideTheta[k3] +
 3 E^(-a k3) k2 k3^2 \[Pi] HeavisideTheta[k3] -
 3 E^(a (k1 - k3))
   k1 k2 (2 + a (k1 - k3)) (k1 - k3) \[Pi] HeavisideTheta[-k1 + k3] +
 3 E^(a (k1 - k3)) k1 (k1 - k3)^2 \[Pi] HeavisideTheta[-k1 + k3] -
 3 E^(a (k1 - k3)) k2 (k1 - k3)^2 \[Pi] HeavisideTheta[-k1 + k3] +
 3 E^(a (k2 - k3))
   k1 k2 (2 + a (k2 - k3)) (k2 - k3) \[Pi] HeavisideTheta[-k2 + k3] +
 3 E^(a (k2 - k3)) k1 (k2 - k3)^2 \[Pi] HeavisideTheta[-k2 + k3] -
 3 E^(a (k2 - k3)) k2 (k2 - k3)^2 \[Pi] HeavisideTheta[-k2 + k3] -
 3 E^(a (k1 + k2 - k3))
   k1 (k1 + k2 - k3)^2 \[Pi] HeavisideTheta[-k1 - k2 + k3] +
 3 E^(a (k1 + k2 - k3))
   k2 (k1 + k2 - k3)^2 \[Pi] HeavisideTheta[-k1 - k2 + k3]
*)
$\endgroup$
1
  • $\begingroup$ This works. Thanks for your "high-level" tricks. $\endgroup$
    – Laplacian
    Feb 20 at 5:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.